Operator Guide

Everything you need to deploy and run an operator on the Tangle Router network.

What is an Operator?

An operator is an independent infrastructure provider that deploys a Tangle Blueprint to serve AI model inference. Operators stake TNT as collateral, set their own pricing, and earn 80% of inference revenue.

Blueprint Types

vLLM Blueprint

Serve open-weight models (Llama, Mistral, Qwen, DeepSeek) on your own GPU hardware using vLLM.

Modal Blueprint

Serverless GPU inference via Modal. Auto-scales, no hardware to manage.

Custom Blueprint

Build your own Blueprint with full control over the inference pipeline. See the Blueprint documentation to get started.

Quick Start

# Clone the vLLM Blueprint
git clone https://github.com/tangle-network/vllm-inference-blueprint
cd vllm-inference-blueprint

# Configure
cp operator/config.example.toml operator/config.toml
# Edit: model, GPU count, pricing, endpoint URL

# Build and run
cargo build --release
./target/release/operator --config operator/config.toml

# Register on Tangle
tangle operator register \
  --blueprint-id <id> \
  --endpoint https://your-operator.example.com \
  --stake 10000

Requirements

  • GPU hardware (NVIDIA A100/H100) or Modal/cloud account
  • 10,000 TNT minimum stake
  • Publicly accessible HTTPS endpoint
  • Health check endpoint at /health
  • <200ms response time from gateway